21 – Determining trends in irruptive desert species^287examples. The plains mouse Pseudomys australis occupies varied habitats when its
numbers increase post-rain, but retreats to localised areas of cracking clay when
conditions dry out (Pavey et al. 2017). Drought-refuges may be fixed in particular
locations in the landscape (Greenville et al. 2013) or shift between locations
following local rainfalls that generate ephemeral pulses of productivity (Newsome
and Corbett 1975). Monitoring in known or likely refuge habitats should allow
more consistent tracking of population trends than monitoring in habitats that are
used only infrequently and after widespread rainfall, but such information is
available at present for very few species (Pavey et al. 2017).
Fourth, we advocate a return to basics; that is, for a restoration of the value
placed on natural history. The three points noted above are all predicated on the
assumption that managers and researchers know enough about the basic biology of
their target organisms to know which monitoring methods are best and when and
where the organisms are active. For target organisms that are poorly known, it is
essential to allocate some initial resources to improve understanding of their basic
ecology and resource needs, and thus ensure that the design of subsequent
monitoring is effective for them.
Fifth, monitoring should not stop at the target threatened species. Although
these taxa are obviously of primary interest, if we are monitoring irruptive desert
species we know already that their numbers will f luctuate – often wildly. To
understand what drives the f luctuations, simultaneous monitoring of
environmental factors and known or putative threats should also be undertaken. If
f luctuations of the target species appear to fall outside previously observed bounds,
or show patterns that are unexpected based on knowledge of the species’ natural
history, analysis of the covariates then may identify the potential cause of the
perturbations. If a given threat appears to be driving populations of the target
species to levels that are deemed unacceptably low, informed management
intervention can be implemented. Precise trigger points for such interventions are
difficult to specify, but obviously need to occur before zero detections are recorded
in one or more monitoring sites: the lower the numbers of a target species, the
greater the chance it will disappear. Species that respond asynchronously, such as
S. youngsoni, are more likely to be inf luenced by factors at a local scale and thus
vulnerable to local extinction events, whereas species that show synchronous
spatial dynamics, such as D. blythi, respond to regional scale events and are at
greater risk of widespread extinction. General examples of detecting trigger points
for management intervention are provided by Caughley and Gunn (1996), and
specific examples for irruptive species by Letnic and Dickman (2006).
Sixth, as monitoring results accrue, it is essential that the data are maintained
safely and securely, cleaned, and made available to relevant stakeholders for review.
This is necessary so that robust analysis and modelling can be undertaken to
uncover trends in the target species and identify factors that may be driving them.